Skip to content

husky: add splash to VENDOR_CMDLINE - #278

Closed
Somnius wants to merge 1 commit into
derekross:mainfrom
Somnius:husky-splash
Closed

Somnius wants to merge 1 commit into
derekross:mainfrom
Somnius:husky-splash

Conversation

@Somnius

@Somnius Somnius commented Nov 1, 2026

Copy link
Copy Markdown
Contributor

@derekross this is the husky-only splash change from #266.

One line in husky's VENDOR_CMDLINE, and the comment above it:

console=tty0 splash plymouth.graphical plymouth.force-scale=1 vt.global_cursor_default=0 clk_ignore_unused pd_ignore_unused printk.devkmsg=on bootconfig

You asked for splash. I also added raven's Plymouth words next to it (plymouth.graphical, plymouth.force-scale=1, vt.global_cursor_default=0), because that is the line I tested the whole day, and I didn't boot with splash alone. If you want only splash I'll cut the rest, but then it's an untested line. I left out raven's quiet, its console font and its regulator_ignore_unused.

What it does: the shared initramfs/init sees splash in /proc/cmdline, writes /splash.bgra.gz to /dev/fb0 right away and sends its own status lines to kmsg instead of tty0. The frame is made at image time by bin/omamobile: the logo from the Omarchy Plymouth theme (/usr/share/plymouth/themes/omarchy/logo.png, from the omarchy-settings 4.0.6 package) centred on #1a1b26 at husky's PANEL_SIZE 1344x2992, 16,084,992 bytes raw and about 19.7 KB gzipped.

Getting the text back: drop splash from the line and reflash, then the init's status lines show on tty0 again. Kernel messages weren't on the 8 Pro's screen before either: /proc/cmdline on my phone ends up with console=null (the bootloader adds it), and console=tty0 from our line is there too, but I haven't checked what that shows on this panel, so I won't promise more than "the init text comes back".

Tested on my 8 Pro, slot B, kernel 7.3.0-rc6-omamobile-push1-g5899fec2bd28: /proc/cmdline has the new words, Plymouth started at 3.36 s and quit at 4.84 s, the display unit was up at 4.47 s and sddm at 4.85 s, and the video shows the Google logo, the Omarchy splash for about 8 s, a few seconds of plain dark navy, then the lock screen.

That is not a test of this line alone. The boot also had three local things on the phone that aren't in this PR: the plymouth-quit retain-splash drop-in (copied by hand from the raven/cheetah package; it'll be its own PR for the husky package), a user portal override, and a compositor background colour. What the screen shows between Plymouth quitting at 4.8 s and Hyprland starting is untested without the drop-in. I can flash this line alone and add that boot here if you want.

Checks I ran: tests/device-overlays passes (the descriptor parses; its dtc part skipped, no dtc here), tests/images-layout skipped for lack of AOSP mkbootimg.

The long black wait after the splash before the lock screen is not from this line, it's #277.

@derekross

Copy link
Copy Markdown
Owner

@Somnius thank you. This is the change we agreed on #266, kept to husky's descriptor, with quiet left out and a clear way to get the init text back. Keeping the Plymouth words is fine with me: you tested that line, and it's the same one raven uses.

One thing CI caught: the husky package carries its own copy of the descriptor (packages/omamobile-device-husky/device.conf, installed as /usr/share/omamobile/device.conf), and its test requires the two to be identical, so "Tests" fails with "device.conf drifted from the descriptor". A descriptor change therefore also means a package release. I'll merge #285 first (0.1.5), then this one with three small additions on top:

  • copy device/husky/device.conf into the package;
  • bump omamobile-device-husky to 0.1.6, which makes the merged title omamobile-device-husky 0.1.6: add splash to VENDOR_CMDLINE;
  • wrap the new comment line in device.conf at 80 columns.

I've checked that with those, the package test passes on top of #285. If you'd rather push them yourself, that works too.

Before I merge, please paste the tools/phone-smoke report from a boot with this line. The boot you described on #285, with 0.1.5 installed and this command line, is the one I'd want. Your other local changes on that boot (the portal override and the compositor colour) are fine, as long as the report says they were there.

Thanks also for tracking the long wait down to the portal and filing it separately. I've answered that on #277.

@derekross derekross closed this in 43223ce Nov 1, 2026
@derekross

Copy link
Copy Markdown
Owner

Landed on main as 43223ce (squashed, authored by you), with the edits I described in my last comment. Thank you.

@Somnius

Somnius commented Nov 3, 2026

Copy link
Copy Markdown
Contributor Author

@derekross here is the tools/phone-smoke report you asked for on this. It's not from the exact boot you described (0.1.5 with the new command line): I didn't keep a smoke report from that boot, and the phone has been updated since. This one is from today's boot (the date in the report is the phone's clock, which has no RTC, so it isn't the real time) with husky package 0.1.7 (built from main) and omamobile-shell 0.2.100. The installed /usr/share/omamobile/device.conf has the splash plymouth.graphical plymouth.force-scale=1 part in VENDOR_CMDLINE, and /proc/cmdline of the running boot has those three parameters. All the rows marked required are ok; the table has one not ok, the new dmesg errors:

phone-smoke: husky

Kernel 7.3.0-rc6-omamobile-push1-g5899fec2bd28, tools at 330a7f2, 2026-11-02T06:26Z.

Check Result Detail
boot (required) ok kernel 7.3.0-rc6-omamobile-push1-g5899fec2bd28, slot b, up 55s
storage guard (required) ok 62 partitions; only userdata and slot B's boot images writable
root (required) ok ext4 on userdata, 207G free
kernel health (required) ok no oops or BUG; 0 WARN, 36 error lines in dmesg (28 known, 8 new)
new dmesg errors not ok N N: brcmf_msgbuf_rxbuf_data_post: No PKTID available !! x3; exynos-clkout exynos-clkout: probe with driver exynos-clkout failed with error N x1; exynos-clkout exynos-clkout: cannot match parent device x1
systemd units ok none failed
USB gadget (required) ok bound to 11210000.usb
display ok card2-DSI-1 connected, 1344x2992
GPU ok render node on panthor
session (required) ok Hyprland and the Omarchy shell running
lock ok PAM file present, the shell can lock (passwordPam true, locked true)
touch ok SALIXP Touchscreen
buttons ok gpio-keys
Wi-Fi ok wlp1s0 on brcmfmac, link up
Bluetooth skip no controller
battery ok 55% Charging, 41 °C
sound cards skip none
sensors skip no sensor proxy
cameras skip no libcamera tools
thermal ok hottest zone 65 °C

Packages: hyprland 0.56.2-5.1, libcamera 0.7.2-4.22, mesa 1:26.2.4-1.4, omacam 0.3.3-1, omamessages 0.2.1-1, omamobile-apps 0.2.2-1, omamobile-device-husky 0.1.7-1, omamobile-firstboot 0.1.2-1, omamobile-keyring 20261026-1, omamobile-mirror 0.1.1-1, omamobile-omarchy 0.1.31-1, omamobile-owl-widgets 0.1.2-1, omamobile-shell 0.2.100-1, omaphone 0.2.2-1, omarchy 4.0.6-1, omatel-agent 0.2.4-1, quickshell 0.3.2.r7.g2b4501a-1

The report names three kinds of new dmesg error: No PKTID available (3 lines, the brcmfmac thing from kernel #156), and two exynos-clkout lines (cannot match parent device, then the probe failing; the journal has -22 for it). That is 5 of the 8 lines the report counts and I don't know what the other 3 are. The same two exynos-clkout lines are also in the journals of the two boots before this one, and I haven't looked into them.

Local things on this boot that are not in main: a local service that sets the slot B successful bit in devinfo, a compositor background colour in looknfeel.lua, a blank cursor theme with a script that puts the real one back, and a Lang key bind in bindings.lua. The kernel is the same push1-g5899fec2bd28 build as before. The portal override from before is removed on this boot, so that's the stock omamobile-shell portal setup.

I haven't filmed this boot, so I still can't say the frame stays up until Hyprland draws.

@Somnius

Somnius commented Nov 3, 2026

Copy link
Copy Markdown
Contributor Author

@derekross a follow-up to my smoke report: I filmed a reboot of the same setup (husky package 0.1.7, omamobile-shell 0.2.100, the boot image with splash plymouth.graphical plymouth.force-scale=1). I stepped through the video at 1 frame per second, and at 4 per second only for the stretch from 40 s to 45 s (the end of the boot), so a short black blink between the splash and the next screen could have been missed:

  • Google logo, then the Omarchy Plymouth splash for about 8 s.
  • After the splash I saw no black frame, but at 1 frame per second that's weak evidence: a plain light blue screen for about 4 s, a notification on it, then the desktop and then the lock screen. I can't say what draws the light blue, so I still can't say that the splash frame stays up until Hyprland draws.
  • That light blue is not the navy my local looknfeel.lua sets (background_color = rgb(1a1b26), still there); on the boots on 1 November that stretch was navy. I haven't looked at why.
  • The desktop (wallpaper and the launcher bar) is on screen for about 0.75 s before the lock screen covers it. I don't know if that's new, I haven't filmed the earlier boots with that in mind.
  • systemd-analyze: kernel 1.9 s plus userspace 3.5 s, graphical target after 3.1 s of userspace; Google logo to lock screen in the video is about 14 s.

The video is 50 s long and I can send it if you want to see it.

robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…#278)

The Pixel 8 Pro's VENDOR_CMDLINE gains splash (with the Plymouth words
from raven's tested line), so the shared init draws the static Omarchy
frame at boot. quiet is left out, so the kernel text stays available
during bring-up; removing splash brings the init text back.

Maintainer edits at merge: the package's copy of the descriptor follows
device/husky/device.conf (its test requires them to match), the comment
line is wrapped, and the package version is 0.1.6.

Closes derekross#278

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants